I have 2 brokers on different machines - say A and B. Broker B statically connects to broker A, using the failover:tcp protocol
I have a subscriber to B that connects using the failover:tcp protocol too. I can send messages from a publisher on A to the subscriber on B. I can also see the network connection from B in A's web interface for the queue. If I restart broker A, I can see that broker B has reconnected, but the publisher on A can't send messages to the subscriber on B. When I look at the web interface for the broker, I can see that the subscriber that A had (the incoming network connection) is no longer there. If I restart the subscriber on B or the broker B, it all starts working. So, is this a valid use for the failover I.e. To allow asynchronous broker restarts. If it's not, how can it be done without have to have restart depdendency in place? I tried setting dynamicOnly to true and I tried the UpdateClusterClient option, but neither helped. I was using version 5.3.0 and upgraded to 5.4.2, but again, this made no difference.. Any help is appreciated! -- View this message in context: http://activemq.2283324.n4.nabble.com/Networked-failover-brokers-don-t-forward-subscriptions-tp3482672p3482672.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.
